The question often arises whether an elite education is worth the cost. There are plenty of studies and stories that quantify the net impact. But these studies typically assume the student attends/graduates.
What if we un-bundle the activity? Broadly: acceptance, education, and networking. We can consider this a type of value chain, and efficient-market theory says it should all add up in the end (-ish).
Imagine a high school senior who gets accepted to, say, Williams or Princeton. The student may now claim this distinction, with minimal cost. A few hundred dollars in application fees, perhaps. Maybe an extra thousand or two in test prep. A clever self-promoter might even position himself as deliberately foregoing college to do something even better (Bill Gates, LeBron James).
Since the student does not have to attend once accepted, the education and networking valuations should be done marginally (ignoring the acceptance value). Can the process of undergraduate learning be that different from the highest to lowest ranked colleges? Does a lower ranked school withhold information? Presumably, the classes are fairly introductory and the ones that are in-depth usually fit into a pathway for a graduate school curriculum. Quality of teacher is not what determines the school's ranking, since publishing is paramount. Further, teaching as a skill may in fact be easier when students are more intelligent; the best teachers may actually add the most value with students of lesser aptitude. I don't want to rehash the debate; let's just stipulate that the variance in the ability of professors to impart facts and thought processes is less than the variance in tuition (Do you learn twelve times as much at a $60k school versus a $5k one?). If this is correct, then the extra value must come from some other factor(s).
If student A relies on his classmates to pull him ahead in the workforce, then A is not likely to be one of the top tier students. Same for student B, student C, and the other non-top tier students. So networking may, in fact, be a drain on top tier students themselves. Take from the few, give to the many; this works when you're behind a veil and don't know which group you are in to start with. So there's probably some value in networking to most students, but not all, and the net effect may even be a wash since the value is being redistributed.
There may be another factor that explains the extra value then. One thought is that the undergraduate degree is a qualifier for graduate school. An elite degree is a near requirement for elite grad school admittance. Now, that's life-changing value.
If this overall analysis is accurate, then what should a rational teenager do? The value chain above suggests two optimal branches. If grad school is in your future, then go all-in and graduate. If it's not, then get an admittance letter to a top school and then get a cheaper education or use that letter to go get a good job. (The symmetry here is that an employer would want the admitted student who sees better than to overpay for education itself.)
As always, I'm sure I haven't thought of every nuance so consider this kindling.